Predicting factors associated with the outbreak in clusters of Influenza-Like- Illness during the seasonal period 2012-2014 in Thailand

Abstract

Effective and timely surveillance for influenza to provide early warning and to predict the outbreak have always been the challenging issue. This study aim to determine the predicting factors associated with the occurrence in clusters of Influenza-Like-Illness and early epidemic. We analyzed the surveillance information during 2012-2014 from influenza surveillance, Clusters Reporting from Event-Based Surveillance and Outbreak Investigation of the Bureau of Epidemiology, Influenza-Like-lllness (ILI) electronic reporting from hospitals network and Laboratory Influenza Sentinel Network, Department of Medical Science. Findings: Pattern of clusters followed the seasonal and majority occured during 2 periods: January-February and June- September with minor variation. Regression analysis identified ILI level > 5.00% and proportion of influenza virus tested positive > 15.00% were significantly and independently associated with the frequency of clusters. The number of clusters increase 2 times when ILI increase from under 4.00% to 7.00% and the proportion of tested positive increase to 15.00-30.00%. The interaction of ILI and proportion of tested positive was observed at the early epidemic. Inter-quartile differences of ILI decrease as the epidemic increased. ILI at 5.00% and proportion of virus positive tested more than 15.00% are optimal early warning signals. ILI is the sensitive indicator for early warning and laboratory surveillance provide specificity. Both variables are useful to monitor the course of epidemic and should be used in national control program and pandemic preparedness.